Hello,
have you checked MoreLikeThis feature?
On Tue, May 15, 2012 at 11:26 PM, Samarendra Pratap wrote:
> - We are calculating frequency of category ids in these top results. We
> are not using facets because that gives count for all, relevant or
> irrelevant, results.
> - Based on cate
just use the admin UI -- look at the 'cloud' tab
On Tue, May 15, 2012 at 12:53 PM, Naga Vijayapuram wrote:
> Alright; thanks. Tried with "-OPTIONS=jsp" and am still seeing this on
> console Š
>
> 2012-05-15 12:47:08.837:INFO:solr:No JSP support. Check that JSP jars are
> in lib/jsp and that th
Rahul,
Get SPM for Solr from http://sematext.com/spm and you'll get all the insight
into your cache utilization you need and more, and through it you will get
(faster) answers to all your questions if you play with your Solr config
settings and observe cache metrics in SPM.
Otis
Performan
Hi Naga,
I'll guess . Fall 2012.
Otis
Performance Monitoring for Solr / ElasticSearch / HBase -
http://sematext.com/spm
>
> From: Naga Vijayapuram
>To: "solr-user@lucene.apache.org"
>Sent: Tuesday, May 15, 2012 5:17 PM
>Subject: - When is Solr 4.0
Hi,
I don't think you can set that, but you may still want to upgrade. Solr 3.6
has a lower memory footprint, is faster, and has more features.
Otis
Performance Monitoring for Solr / ElasticSearch / HBase -
http://sematext.com/spm
>
> From: Jon Kirton
Erick
Yes thanks I did see that and am working on a solution to that already.
Hope to post a new revision shortly and eventually migrate to the extension
"store".
Cheers
Amit
On May 15, 2012 9:20 AM, "Erick Erickson" wrote:
> I think I put one up already, but in case I messed up github, complex
hello
thanks for the reply
this is the output - docsPending = 0
commits : 1786
autocommit maxDocs : 1000
autocommit maxTime : 6ms
autocommits : 1786
optimizes : 3
rollbacks : 0
expungeDeletes : 0
docsPending : 0
adds : 0
deletesById : 0
deletesByQuery : 0
errors : 0
cumulative_adds : 178775
OK, setting the wait_timeout back to its previous value and adding readOnly
didn't help, I got the stack overflow again. I re-upped the mysql timeout
value again.
-jsd-
On Tue, May 15, 2012 at 2:42 PM, Jon Drukman wrote:
> I fixed it for now by upping the wait_timeout on the mysql server.
>
Hi,
Would distributed search (the old way where you provide the solr host
IP's etc.) still work between different solrclouds?
thanks,
Darren
Just tested to make sure. queryNorm is changing after you add bq parameter. For
example : 0.00317763 = queryNorm becomes 0.0028020076 = queryNorm.
Since all scores are multiplied by this queryNorm factor, score of a document (
even if it is not effected/boosted by bq) changes.
before
Hi Samarendra,
This does look like a candidate for a custom query component if you want to do
this inside Solr. You can of course continue to do this at the client.
-sujit
On May 15, 2012, at 12:26 PM, Samarendra Pratap wrote:
> Hi,
> I need a suggestion for improving relevance of search resul
We're running solr v1.4.1 w/ approx 30M - 40M records at any given time.
Often, socket timeout exceptions occur for a search query. Is there a
compelling reason to upgrade? I.e. can u set a socket timeout in
solrconfig.xml in the latest version and not in v1.4.1 ?
I fixed it for now by upping the wait_timeout on the mysql server.
Apparently Solr doesn't like having its connection yanked out from under
it and/or isn't smart enough to reconnect if the server goes away. I'll
set it back the way it was and try your readOnly option.
Is there an option with Dat
> If the bq is only supposed apply the
> boost when the field value is greater
> than 0.01 why would trying another query make sure this is
> working.
>
> Its applying the boost to all the fields, yes when the boost
> is high enough
> most of documents with a value GT 0.01 show up first however
>
Shot in the dark here, but try adding readOnly="true" to your dataSource tag.
This sets autocommit to true and sets the Holdability to
ResultSet.CLOSE_CURSORS_AT_COMMIT. DIH does not explicitly close resultsets
and maybe if your JDBC driver also manages this poorly you could end up with
stra
… Any idea, anyone?
Thanks
Naga
Finally got a handle on this by looking into the New Admin UI -
http://localhost:8983/solr/#/~cloud
Thanks
Naga
On 5/15/12 12:53 PM, "Naga Vijayapuram" wrote:
>Alright; thanks. Tried with "-OPTIONS=jsp" and am still seeing this on
>console Š
>
>2012-05-15 12:47:08.837:INFO:solr:No JSP support
i don't think so, my config is straightforward:
i'm triggering the import with:
http://localhost:8983/solr/dataimport?command=full-import&clean=true&commit=true
On Tue, May 15, 2012 at 2:07 PM, Michael Della Bitta <
michael.della.bi...@appinions.com
Hello, geeky2:
In statistics in the update section, do you see a non-zero value for
docsPending?
Thanks,
Michael
On Tue, May 15, 2012 at 4:49 PM, geeky2 wrote:
>
> hello,
>
> After doing a DIH full-import (with clean=true) after deleting records in
> the database, i noticed that the number of
Hi, Jon:
Well, you don't see that every day!
Is it possible that you have something weird going on in your DDL
and/or queries, like a tree schema that now suddenly has a cyclical
reference?
Michael
On Tue, May 15, 2012 at 4:33 PM, Jon Drukman wrote:
> I have a machine which does a full update
If the bq is only supposed apply the boost when the field value is greater
than 0.01 why would trying another query make sure this is working.
Its applying the boost to all the fields, yes when the boost is high enough
most of documents with a value GT 0.01 show up first however since it is
applyi
> Scratch that...it still seems to be
> boosting documents where the value of
> the field is empty.
>
>
> bq=regularprice:[0.01 TO *]^50
>
> Results with bq set:
>
>
> name="score">2.2172112
> name="code">bhl-ltab-30
>
>
>
> Results without bq set:
>
>
> name="score">2.48
hello,
After doing a DIH full-import (with clean=true) after deleting records in
the database, i noticed that the number of documents processed, did change.
example:
Indexing completed. Added/Updated: 595908 documents. Deleted 0 documents.
however, i noticed the numbers on the statistics page
Hello Neil,
if "manipulating tf" is a possible approach, why don't extend
KeywordTokenizer to make it work in the following manner:
"3|wheel" -> {wheel,wheel,wheel}
it will allow supply your per-term-per-doc boosts as a prefixes for field
values and multiply them during indexing internally.
The
Thanks
--
View this message in context:
http://lucene.472066.n3.nabble.com/Show-a-portion-of-searchable-text-in-Solr-tp3983613p3983942.html
Sent from the Solr - User mailing list archive at Nabble.com.
On Tue, May 15, 2012 at 2:47 PM, Naomi Dushay wrote:
> We are using the ICUFoldingFilterFactory with great success to fold
> diacritics so searches with and without the diacritics get the same results.
>
> We recently discovered we have some Korean records that use an alif diacritic
> instead of
Hello Hoss and the list,
We are currently using Lucene payloads to store per-document-per-keyword
scores for our dataset. Our dataset consists of photos with keywords
assigned (only once each) to them. The index is about 90 GB, running on
24-core machines with dedicated 10k SAS drives, and 16/32 G
Alright; thanks. Tried with "-OPTIONS=jsp" and am still seeing this on
console Š
2012-05-15 12:47:08.837:INFO:solr:No JSP support. Check that JSP jars are
in lib/jsp and that the JSP option has been specified to start.jar
I am trying to go after
http://localhost:8983/solr/collection1/admin/zook
Scratch that...it still seems to be boosting documents where the value of
the field is empty.
bq=regularprice:[0.01 TO *]^50
Results with bq set:
2.2172112
bhl-ltab-30
Results without bq set:
2.4847748
bhl-ltab-30
On Tue, May 15, 2012 at 12:40 PM, Donald Organ wrote
Hello,
I am trying to understand how I can size the caches for my solr powered
application. Some details on the index and application :
Solr Version : 1.3
JDK : 1.5.0_14 32 bit
OS : Solaris 10
App Server : Weblogic 10 MP1
Number of documents : 1 million
Total number of fields : 1000 (750 strings, 2
I have already triple cross-checked that all my clients are using
same version as the server which is 3.6
Thanks
Ravi Kiran
On Tue, May 15, 2012 at 2:09 PM, Ramesh K Balasubramanian
wrote:
> I have seen similar errors before when the solr version and solrj version in
> the client don't match.
We are using the ICUFoldingFilterFactory with great success to fold diacritics
so searches with and without the diacritics get the same results.
We recently discovered we have some Korean records that use an alif diacritic
instead of an apostrophe, and this diacritic is NOT getting folded. Has
I have seen similar errors before when the solr version and solrj version in
the client don't match.
Best Regards,
Ramesh
Thanks, Jack! I think you are right. But I also copied cr_firstname to text,
I assumed Solr would highlight cr_firstname if there is a match. I guess the
only solution is to copy all field to another field which is not tokenized.
Yes, it is "firstname", good catch!
Thanks again!
TJ
--
View this
Hello,
Unfortunately it seems like I spoke too early. Today morning I
received the same error again even after disabling the iptables. The
weird thing is only one out of 6 or 7 queries fails as evidenced in
the stack traces below. The query below the stack trace gave a
'status=500' subsequen
That is the default response format. If you would like to change that, you
could extend the search handler or post process the XML data. Another option
would be to use the javabin (if your app is java based) and build xml the way
your app would need.
Best Regards,
Ramesh
>
In 4.0, solr no longer uses JSP, so it is not enabled in the example setup.
You can enable JSP in your servlet container using whatever method
they provide. For Jetty, using start.jar, you need to add the command
line: java -jar start.jar -OPTIONS=jsp
ryan
On Mon, May 14, 2012 at 2:34 PM, Nag
I have figured it out using your recommendation...I just had to give it a
high enough boost.
BTW its a float field
On Tue, May 15, 2012 at 9:21 AM, Ahmet Arslan wrote:
> > The problem with what you provided is
> > it is boosting ALL documents whether
> > the field is empty or not
>
> Then a
Hi,
I have a few questions, please bear with me:
1- I have a theory. nutch may be used to index to solr when we don't
have access to URL's file system, while we can use curl when we do have
access. Am I correct?
2- A tutorial I have been reading is talking about different levels of
id. Is the
In the case of text:"G-Money", the term is analyzed by Solr into the phrase
"g money", which matches in the text field, but will not match for a string
field containing the literal text "G-Money". But when you query
cr_fristname:"G-Money", the term is not tokenized by the Solr analyzer
because
I am also working on highlighting. I don't think so. And the ids in the
highlighting part are the ids of the docs retrieved.
--
View this message in context:
http://lucene.472066.n3.nabble.com/Highlight-feature-tp3983875p3983887.html
Sent from the Solr - User mailing list archive at Nabble.com.
Hello,
>From the response you pasted here, it looks like the field
"itemNoExactMatchStr"
never matched.
Can you try matching in that field only and ensure you have matches ? Given
the ^30 boost, you should have high scores on this field...
Hope this helps,
--
Tanguy
2012/5/15 geeky2
> Hello a
Right, but for that to work, there's an implicit connection between
the docs in core1 and core0, I assume provided by 123456 as
a foreign key or something. There's nothing automatically built
in like this in Solr 1.4 (joins come close, but those are trunk).
Whenever you try to make Solr act just l
Hi Jack,
Thanks for your reply. I did not specify dismax when query with highlighting
enabled: q=text:"G-Money"&hl=true&hl.fl=*, that was the whole query string I
sent. What puzzled me is that the "string" field "cr_firstname" was copied
to text, but it was not highlighted. But if I use
q=cr_frist
Hello all,
i am trying to tune our core for exact matches on a single field (itemNo)
and having issues getting it to work.
in addition - i need help understanding the output from debugQuery=on where
it presents the scoring.
my goal is to get exact matches to arrive at the top of the results.
Hi Erick ,
My Schema is as follows
My data which i am indexing in core0 is
id:1, value:'123456', column1:'A',column2:'null'
id:2, value:'1234567895252', column1:'B',column2:'null'
My data which i am indexing in core1 is
id:3, value:'123456', column1:'null', colum
Hello friends
I have noticed that the highlighted term of a query are returned in a second
xml struct(named "highlighting"). Is it possible to return the highlighted
terms into the doc field. I don't need the solr generated ids of the
highlighted field.
Thanks,
Tom
--
View this message in contex
I think I put one up already, but in case I messed up github, complex
params like the fq here:
http://localhost:8983/solr/select?q=:&fq={!geofilt sfield=store
pt=52.67,7.30 d=5}
aren't properly handled.
But I'm already using it occasionally
Erick
On Tue, May 15, 2012 at 10:02 AM, Amit Nith
SOLR-3430 is already applied to the latest 3.6 and 4.x (trunk) source code. Be
sure you have sources from May 7, 2012 or later (for 3.6 this is SVN r1335205 +
; for trunk it is SVN r1335196 + ) No patches are needed.
About the "modern compiler" error, make sure you're running a 1.6 or 1.7 JDK
Let's go back to this step where things look correct, but we ran into the
edismax bug which requires that you put a space between each left
parenthesis and field name.
First, verify that you are using edismax or not.
Then, change:
&q=chairs AND (regularprice:*^5 OR (*:* -regularprice:*)^0.5)&
Jan
Thanks for your feedback! If possible can you file these requests on the
github page for the extension so I can work on them? They sound like great
ideas and I'll try to incorporate all of them in future releases.
Thanks
Amit
On May 11, 2012 9:57 AM, "Jan Høydahl" wrote:
> I've been testing
> > The problem with what you
> provided is
> > it is boosting ALL documents whether
> > the field is empty or not
>
> Then all of your fields are non-empty? What is the type of
> your field?
How do you feed your documents to solr? My be you are indexing empty string? Is
your field indexed="
oh yeah, forgot about negatives and *:*...
thanks
--
View this message in context:
http://lucene.472066.n3.nabble.com/adding-an-OR-to-a-fq-makes-some-doc-that-matched-not-match-anymore-tp3983775p3983863.html
Sent from the Solr - User mailing list archive at Nabble.com.
Thank you :)
Gary
Le 15/05/2012 15:27, Jack Krupansky a écrit :
Solr is probably simply using the Java JVM default. Set the
java.io.tmpdir system property. Something equivalent to the following:
java -Djava.io.tmpdir=/mytempdir ...
On Windows you can set the TMP environment variable.
-- Jac
Solr is probably simply using the Java JVM default. Set the java.io.tmpdir
system property. Something equivalent to the following:
java -Djava.io.tmpdir=/mytempdir ...
On Windows you can set the TMP environment variable.
-- Jack Krupansky
-Original Message-
From: G.Long
Sent: Tuesda
> The problem with what you provided is
> it is boosting ALL documents whether
> the field is empty or not
Then all of your fields are non-empty? What is the type of your field?
The problem with what you provided is it is boosting ALL documents whether
the field is empty or not
On Tue, May 15, 2012 at 3:52 AM, Ahmet Arslan wrote:
> > Basically I want documents that have a given field populated
> > to have a
> > higher score than the documents that dont. So if you
>
See the "/browse" request handler in the example config.
Only stored fields will be highlighted.
-- Jack Krupansky
-Original Message-
From: Shameema Umer
Sent: Tuesday, May 15, 2012 2:59 AM
To: solr-user@lucene.apache.org
Subject: Re: Show a portion of searchable text in Solr
Can
Hi :)
I'm using SolrJ to index documents. I noticed that during the indexing
process, .tmp files are created in my /tmp folder. These files contain
the xml commands for the documents I add to the index.
Can I change this folder in Solr config and where is it?
Thanks,
Gary
Yes, the parentheses are needed to prioritize the operator precedence (do
the ANDs and then OR those results.) And, add a space after both left
parentheses to account for the edismax bug.
(https://issues.apache.org/jira/browse/SOLR-3377)
-- Jack Krupansky
-Original Message-
From: And
By removing the defType you reverted to using the traditional Solr/Lucene
query parser which supports the particular query syntax you used (as long as
"AND" is in upper-case) and without the parenthesis bug of edismax.
-- Jack Krupansky
-Original Message-
From: Peter Kirk
Sent: Tuesd
> But whether or not it was simply the act of restarting - I'm
> not sure. (I had also found out that "AND " and "OR" should
> be written in uppercase, but this made no difference until
> after I had restarted).
By the way, there is a control parameter for this.
"lowercaseOperators A Boolean
Are you using the edismax query parser (which permits lower case "and" and
"or" operators)? If so, there is a bug with parenthesized sub-queries. If
you have a left parenthesis immediately before a field name (which you do in
this case) the query fails. The short-term workaround is to place a sp
Hi
If I understand the terms correctly, the search-handler was configured to use
"edismax".
The start of the configuration in the solrconfig.xml looks like this:
edismax
In any case, when I commented-out the "deftype" entry, and restarted the solr
webapp, things began to function
Afaik we disabled JSP-Functionality in SOLR-3159 while upgrading Jetty ..
On Tuesday, May 15, 2012 at 1:44 PM, Erick Erickson wrote:
> What do you mean "jsp support"? What is it you're trying to do
> with jsp? What servelet container are you using? Details matter.
>
> Best
> Erick
>
> On Mon
I really don't understand what you're trying to achieve.
"query : column1:"A" should searched in core0 and column2:"B" should be
searched in core1 and later the results from both queries should use
condition AND and give final response.?"
core1 and core0 are completely separate cores, with separa
Hi,
it is AND (uppercase) not and (smallcase) (and OR instead of or).
Regards,
Peter
2012/5/15 András Bártházi :
> Hi,
>
> You should use parantheses, have you tried that?
> q=(skcode:2021051 and flength:368.0) or (skcode:2021049 and
> ent_no:1040970907)
>
> http://robotlibrarian.billdueber.com/
Hi,
We have checked out the latest version of Solr source code from svn. We are
trying to apply the following patch file to it.
https://issues.apache.org/jira/browse/SOLR-3430
While applying the patch file using eclipse (i.e. using team-->apply patch
options), we are getting cross marks for ce
Yes. In fact, all the caches get flushed on every commit/replication cycle.
Some of the caches get autowarmed when a new searcher is opened,
which happens...you guessed it...every time a commit/replication happens.
Best
Erick
On Tue, May 15, 2012 at 1:32 AM, shinkanze wrote:
> hi ,
>
> I want
What do you mean "jsp support"? What is it you're trying to do
with jsp? What servelet container are you using? Details matter.
Best
Erick
On Mon, May 14, 2012 at 5:34 PM, Naga Vijayapuram wrote:
> Hello,
>
> How do I enable JSP support in Solr 4.0 ?
>
> Thanks
> Naga
> It doesn't make a difference. But now I'm thinking there's
> something completely odd - and I wonder if it's necessary to
> use a special search-handler to achieve what I want.
>
> For example, if I execute
> q=(skcode:2021051 AND flength:368.0)
>
> I get no results. If I omit the parentheses
Hi - thanks for the response. Yes I have tried with parentheses, to group as
you suggest.
It doesn't make a difference. But now I'm thinking there's something completely
odd - and I wonder if it's necessary to use a special search-handler to achieve
what I want.
For example, if I execute
q=(
Hi,
You should use parantheses, have you tried that?
q=(skcode:2021051 and flength:368.0) or (skcode:2021049 and
ent_no:1040970907)
http://robotlibrarian.billdueber.com/solr-and-boolean-operators/
Bye,
Andras
2012/5/15 Peter Kirk
> Hi
>
> Can someone please give me some help with a simple q
> that does not change the results for
> me:
>
> -suggest?q=suggest_terms:lap*&fq=type:P&fq=((-type:B))&debugQuery=true
> -found 1
>
> -suggest?q=suggest_terms:lap*&fq=type:P&fq=((-type:B)+OR+name:aa)&debugQuery=true
> -found 0
>
Negative clause and OR clause does not work like this.
fq=+*:* -t
that does not change the results for me:
-suggest?q=suggest_terms:lap*&fq=type:P&fq=((-type:B))&debugQuery=true
-found 1
-suggest?q=suggest_terms:lap*&fq=type:P&fq=((-type:B)+OR+name:aa)&debugQuery=true
-found 0
looks like a bug?
xab
--
View this message in context:
http://lucene.472066.n3.nab
Hi,
You can't. Try eDisMax instead: http://wiki.apache.org/solr/ExtendedDisMax
--
Jan Høydahl, search solution architect
Cominvent AS - www.facebook.com/Cominvent
Solr Training - www.solrtraining.com
On 15. mai 2012, at 11:05, ayyappan wrote:
> Hi
>
> My queries are working with standard
I have written an article on this. The various steps to restrict /
authenticate Solr admin interface.
http://www.findbestopensource.com/article-detail/restrict-solr-admin-access
Regards
Aditya
www.findbestopensource.com
On Thu, Mar 29, 2012 at 1:06 AM, geeky2 wrote:
> update -
>
> ok - i was
Hi
My queries are working with standard query handler but not in dismax.
*it is working fine *
EX :
q=scanneddate:["2012-02-02T01:30:52Z" TO "2011-09-22T22:40:30Z"] .
*Not Working :*
EX
defType=dismax&q=["2012-02-02T01:30:52Z" TO
"2011-09-22T22:40:30Z"]&qf=scanneddate
How can I check for
Without reading the whole thread let me say that you should not trust
the solr admin analysis. It takes the whole multiword search and runs
it all together at once through each analyzer step (factory).
But this is not how the real system works. First pitfall, the query parser
is also splitting at w
Hi
Can someone please give me some help with a simple query.
If I search
q=skcode:2021051 and flength:368.0
I get 1 document returned (doc A)
If I search
q=skcode:2021049 and ent_no:1040970907
I get 1 document returned (doc B)
But if I search
q=skcode:2021051 and flength:368.0 or skcode:2021
HI,
I want to configured 2 cores in my SOLR instance. Now i want to query core0
with different query and core1 with diffrent query and finally merge the
results .
Please suggest me the best way to do this .
Thanks
Ravi
--
View this message in context:
http://lucene.472066.n3.nabble.com/Query-
> Basically I want documents that have a given field populated
> to have a
> higher score than the documents that dont. So if you
> search for foo I want
> documents that contain foo, but i want the documents that
> have field a
> populated to have a higher score...
Hi Donald,
Since you are usi
Thanks for the suggestions.
I tried to use SolrJ within my Servlet. Although the SolrJ QueryResponse is
not returning a well formed Json Object.
I need the Json String with quotes as below. although
QueryResponse.toString() doesn't return json with quotes at all.
jsonp1337064466204({"responseHea
if i use
q=scanneddate:["2011-09-22T22:40:30Z" TO "2012-02-02T01:30:52Z"] .
it is working fine .
but when i tried with dismax query .it is not working .
EX :
select/?defType=dismax&q=["2011-09-22T22:40:30Z" TO
"2012-02-02T01:30:52Z"]&qf=scanneddate&version=2.2&start=0&rows=50&indent=on&wt=json&&de
Thanks Tommaso .
Could you please tell me is their any way to get this scenario to get
worked?
http://localhost:8983/solr/core0/select?shards=localhost:8983/solr/core0,localhost:8983/solr/core1&q=column1:"A";
AND column2:"B"
Is their any way we can achieve this below scenario
query : column1:
Can somebody tell me where should I place the highlighting parameters, when
I did on the query, it is not working.
&hl=true&hl.requireFieldMatch=true&hl.fl=*
FYI: I am new to solr. My aim is to have emphasis tags on the queried
words and need to display only the query relevant snippet of the
Th
86 matches
Mail list logo
